Key Points
- Lightricks’ AI video model runs locally on consumer devices.
- Co‑developed with Nvidia and debuted at CES 2026.
- Generates up to 20‑second clips at 50 fps with 4K resolution and native audio.
- Available as an open‑weight model on HuggingFace and ComfyUI.
- On‑device processing gives creators full data control and faster generation (1‑2 minutes per prompt).
- Reduces reliance on cloud services, lowering cost and protecting intellectual property.
On‑Device AI Video Innovation
Lightricks announced a new AI video model that can operate directly on users’ devices, a capability that sets it apart from most AI video generators that depend on cloud‑based data centers. The model, co‑developed with Nvidia, was unveiled at CES 2026, where Nvidia also highlighted a suite of AI‑enhanced tools for gamers and other applications.
Technical Capabilities
The model is designed to create video clips up to 20 seconds long at a frame rate of 50 frames per second. It supports 4K resolution and includes native audio, positioning it for professional‑grade creative projects. Generation times are typically 1‑2 minutes per prompt, offering faster turnaround compared with cloud‑only solutions.
Open‑Weight Release
Lightricks made the model available as an open‑weight model on platforms such as HuggingFace and ComfyUI. While the model is not fully open‑source, the open‑weight approach provides developers with insight into the model’s architecture without revealing every detail.
Benefits of Local Execution
Running the model on a device gives creators full control over their data, eliminating the need to share content with large tech firms that might use it to improve their own AI systems. This privacy advantage is particularly important for entertainment studios that must protect intellectual property. Additionally, local processing can reduce costs and save time, as the average prompt generates a video in 1‑2 minutes.
Implications for Creators
The combination of high‑quality output, on‑device performance, and open‑weight accessibility makes the Lightricks model an attractive option for a range of creators, from individual filmmakers to major studios. Its ability to deliver professional‑grade video without relying on external servers addresses both security and efficiency concerns in the growing field of generative AI.
Source: cnet.com